advanced · self-paced
Profiling and JVM Performance Tuning
GC logs, async profilers, and allocation forensics you can repeat after the course ends.
6 weeks · 8h/week
JPY 132,000
Informational rate—final fees appear on your enrollment agreement.
Request information Money-back policyProgram narrative
Performance work is presented as detective stories. You will read GC logs, compare G1 versus ZGC trade-offs for sample workloads, and interpret flame graphs without drowning in micro-optimizations. Includes a capstone where you present findings to a mentor acting as skeptical engineering manager.
What ships in the box
- jcmd and Async Profiler labs on provided samples
- GC log parsing workshop with GCViewer and manual reading
- Heap dump post-mortem with MAT walkthrough
- JFR recording templates for staging and production
- Checklist for safe profiling in production
- Storytelling coaching for performance review meetings
- Capstone rubric aligned to hiring panel questions
Outcomes you can describe
- Narrate a performance investigation with evidence slides
- Choose GC settings with explicit workload assumptions
- Decline premature optimization with diplomatic data
Lead mentor
Mika Okada
Lead instructor; former platform engineer shipping JVM services in Osaka.
Cohort questions
Labs run on Linux containers; macOS and Windows hosts supported with notes.
Profiling labs benefit from 16GB RAM; smaller machines can use cloud workspace option.
We do not tune bare-metal network cards or kernel bypass frameworks.
Experience notes
JFR template landed in our runbooks—on-call thanked me quietly.
Manager role-play was awkward and useful—trimmed three slides of hand-waving.